This kind of relationship, where the user trades information about him- or herself for service is sadly all too common on today's internet, and amounts to voluntary participation in a vast spyware network, where you are ever more effectively watched and tracked no matter what you do.
Technologically, it should be quite feasible to do all the RSS feed gathering and filtering for this sort of service on the client's machine, using software installed locally on the user's machine rather than on the service's servers.
Another possibility is to do the processing on the service's servers, but to try to build in some sort of privacy and anonymity guarantees in to the service itself. This is inherently more problematic, since it would really require some sort of verification of those guarantees by a neutral, trusted third-party before it could be completely trusted. But it would still be a big step forward over what we have now.
Good luck with your business, but I'm not going to use it until and unless I can feel secure that I'm not giving you any information about myself, my tastes or my preferences by using it.